Empowering Digital Entertainment in Georgia with AI-driven Customer Support 

Play videoPlay video

Crystalbet

Gaming

Visit SiteVisit Site

Founded in 2016, Crystalbet offers a broad portfolio of betting, casino, and gaming products through its website and mobile app.

Industry

Gaming

Market

Georgia

Company size

250-500 employees

Monthly active users (chat volume)

Tens of thousands of active players and support conversations

Septembet 17, 2025/4 min read

Share

The challenge

During high-volume periods, such as major sporting events, high-profile tournaments, or promotional campaigns, Crystalbet experiences sharp surges in customer inquiries. These typically include questions related to deposits and withdrawals, account access, verification, bonus terms, and bet settlement.
Without automation or scalable tooling, support teams face increased operational pressure: queues form quickly, response times increase, and maintaining consistent service quality becomes more challenging during peak demand.

The solution: AI agent implementation

To address these challenges, Crystalbet adopted an AI agent (chatbot / virtual assistant) integrated directly into its customer service channels. The AI agent automatically handles routine and frequent customer queries, while seamlessly escalating more complex or sensitive issues to human operators.

Features of the system include:

• A structured knowledge base covering FAQs, account management, payments, bonuses, and game rules
• Smart fallback escalation when the AI cannot confidently resolve a request
• Multichannel access across web and mobile platforms
• Monitoring of key indicators such as automated resolution coverage, first response behaviour, and escalated tickets

With Briano's support, we are able to handle peak traffic far more efficiently. Customers get answers right away, while our teams can focus on complex cases that truly require human attention.

Juna Tatarishvili

Crystalbet

The results

In terms of scale, the impact is clear. Crystalbet serves tens of thousands of monthly active players, with support receiving tens of thousands of chat requests per month. During high-demand periods such as major sports events or promotional campaigns, incoming chat volume can increase several times over the usual level.

Based on the most recent month of observations, the AI agent now independently resolves approximately 45-50% of all incoming chat requests end-to-end. These include routine questions related to payments, account access, bonus instructions, verification guidance, and general site navigation.

Smart escalation efficiency

For the remaining 50-55% of conversations that require human involvement, the AI continues to play a critical supporting role. It provides the first response in 100% of conversations, immediately engaging users and reducing operator workload from the first second. In addition, conversations are forwarded to operators together with a structured summary of the user's latest request, enabling faster and more accurate responses.

While detailed waiting-time dashboards are still being finalised due to the rollout of a new call-centre application, observed operator workload already indicates a substantial reduction in queue formation, particularly during peak hours. The combination of near-half full automation and universal first-response automation significantly eases pressure on human agents.

As a result, customer support operations are able to scale more effectively during traffic spikes. Human agents focus on complex or sensitive cases, while the AI efficiently manages the bulk of repetitive and routine questions. This creates a faster, more resilient support system that protects Crystalbet's reputation during its busiest periods and supports customer loyalty year-round.

After adopting the AI agent, Crystalbet observed clear improvements across several key areas of customer service. Response times, which during peak events could previously stretch to several minutes, are now consistently faster due to 100% AI-driven first response across all conversations. Approximately 45–50% of incoming requests are fully resolved by the AI without human intervention, reducing overall manual handling requirements.

While detailed CSAT dashboards are still being finalized, customer satisfaction remains consistently high, including during traffic surges. Operator workload indicators also show clear improvement: automation has led to a significant reduction in queue formation and waiting pressure, particularly during peak demand.

Conclusion

By integrating an AI agent into its customer support stack, Crystalbet is able to better serve its growing user base, especially during high-profile sporting events, promotional days, and other periods of elevated demand. This enables the platform to maintain high service quality, speed, and consistency, while keeping customer satisfaction high and operational pressure under control.

With continued tuning, improved analytics visibility, and strong alignment between AI and human teams, Crystalbet can expect even higher automated resolution rates, lower cost per ticket, and stronger customer loyalty over time.